The latest articles on Forem by Vincent Botta (@0asa). https://forem.com/0asa https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F92353%2F9e5d512b-831c-4c40-927f-6cf21d7b0fce.png https://forem.com/0asa en Vincent Botta Fri, 31 Oct 2025 16:33:55 +0000 https://forem.com/0asa/whisper-menu-bar-3fd4 https://forem.com/0asa/whisper-menu-bar-3fd4 <p>A minimal, clean speech-to-text menu bar application for macOS using OpenAI's Whisper. You can download the script <a href="https://gist.github.com/0asa/3861da03d1f49106e8073416b39a32bd" rel="noopener noreferrer">here</a>.</p> <h3> Features </h3> <ul> <li>🎤 <strong>Push-to-talk</strong>: Hold Option key to record, release to transcribe</li> <li>📋 <strong>Auto-clipboard</strong>: Transcribed text automatically copied to clipboard</li> <li>🔄 <strong>Model selection</strong>: Switch between tiny, base, small, and medium models</li> <li>🎯 <strong>Clean &amp; minimal</strong>: Simple interface, ~300 lines of code</li> </ul> <h3> Requirements </h3> <ul> <li>macOS (tested on macOS 10.15+)</li> <li>Python 3.8 or higher</li> <li>Microphone access permissions</li> </ul> <h3> Installation and Usage </h3> <ol> <li><p>Copy the <code>whisper-push-to-talk.py</code> somewhere</p></li> <li><p>Install <a href="https://docs.astral.sh/uv/" rel="noopener noreferrer">uv</a> and run:<br> </p></li> </ol> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code> uv run whisper-push-to-talk.py </code></pre> </div> <p><strong>Note</strong>: On macOS, you may need to install PortAudio first for PyAudio:<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code> brew <span class="nb">install </span>portaudio </code></pre> </div> <ol> <li>Grant microphone permissions to Terminal/your Python app when prompted</li> </ol> <p>The app will:</p> <ol> <li>Show a microphone icon (🎤) in your menu bar</li> <li>Load the Whisper model in the background (first run may take a moment)</li> <li>Display "Ready" when ready to use</li> </ol> <h4> Recording </h4> <ul> <li> <strong>Option key</strong>: Hold to record, release to transcribe, <strong>Text appears in clipboard</strong> → Ready to paste anywhere</li> <li>Click the menu bar icon → Model → Select your preferred model: <ul> <li> <strong>tiny</strong>: Fastest, lowest accuracy (~1GB)</li> <li> <strong>base</strong>: Good balance (default, ~1GB)</li> <li> <strong>small</strong>: Better accuracy (~2GB)</li> <li> <strong>medium</strong>: Best accuracy (~5GB)</li> </ul> </li> </ul> <h4> Quit </h4> <p>Click the menu bar icon → Quit</p> <h3> License </h3> <p>MIT License - Feel free to modify and distribute</p> <h3> Credits </h3> <p>Built with OpenAI's Whisper model for speech recognition.</p> transcription whisper python macos